Games and Puzzles - Ludus-Opuscularmm.ludus-opuscula.org/PDF_Files/Singmaster_Vanish_10_21(1_2014... · 12 Vanishing Area Puzzles Figure 3: Hooper. Let us look at the classic 8 8

Games and Puzzles

Vanishing Area Puzzles

David SingmasterLondon South Bank University

[email protected]

In Memoriam Martin Gardner

Abstract: Martin Gardner was very fond of vanishing area puzzles and devotedtwo chapters to them in his first book. There are actually two distinct types.Sam Loyd’s Vanishing Chinaman and similar puzzles have pictures which arereassembled so a part of the picture appears to disappear. But the physical arearemains fixed. The second type cuts up an area and reassembles it to producemore or less area, as in the classic chessboard dissection which converts the 8×8square into a 5× 13 rectangle. Gardner had managed to trace such puzzles backto Hooper in 1774. In 1989, I was visiting Leipzig and reading Schwenter whichreferred to an error of Serlio, in his book of 1535. Serlio hadn’t realised that hisdissection and reassembly gained area, but it is clear and this seems to be theorigin of the idea. I will describe the history and some other versions of the idea.

Key-words: Sam Loyd, vanishing area puzzles.

One of Sam Loyd’s most famous puzzles is “The Vanishing Chinaman” or “Getoff the Earth”. Martin Gardner discussed this extensively in [7] and [8].

Figure 1: “Get off the Earth”.

Recrational Mathematics Magazine, Number 1, pp. 10–21

David Singmaster 11

However, the term “vanishing area puzzle” is used for two different types ofpuzzle. In “The Vanishing Chinaman”, no actual area vanishes - it is one of thefigures in the picture that vanishes, so perhaps we should call this a “vanishingobject puzzle”. In a true “vanishing area puzzle”, an area is cut into severalpieces and reassembled to make an area which appears to be larger or smallerthan the original area. These are considerably older than vanishing objectpuzzles. Martin was fond of these puzzles - indeed, his first puzzle book [6]devotes two chapters to such puzzles - still the best general survey of them -and he also wrote several columns about them [7, 8]. The most famous versionof these is the “Checkerboard Paradox” where an 8× 8 checkerboard is cut intofour parts and reassembled into a 5×13 rectangle, with a net gain of one unit ofarea. This article is primarily concerned with the early history of such puzzles.

Figure 2: Schlomilch.

Gardner describes some 18 & 19C versions of this puzzle idea, going back toHooper (1774) [9], and a surprising connection with Fibonacci numbers discov-ered in 1877 [4]. In 1988, I was visiting Leipzig and looked at some obscurebooks in their library and discovered references going back to 1537.


12 Vanishing Area Puzzles

Figure 3: Hooper.

Let us look at the classic 8× 8 to 5× 13 of Fig. 2. The history of this particularversion is obscure. It is shown in Loyd’s Cyclopedia [10] (p. 288 & 378).


David Singmaster 13

Figure 4: LoydCyc288.

For a long time, I tended to ignore this as it seemed smudged. I later saw thathe shows that both rectangles have chessboard colouring, and he is the first toindicate this. But when I went to scan Fig. 5, I realised that the smudge is



Figure 5: LoydCyc278.

deliberate to conceal the fact that the colouring does NOT match up! Indeed,the corners of a 5 × 13 board should all be the same colour, but two of themin the solution arise from adjacent corners of the 8 × 8 chessboard and haveopposite colours!

Loyd also poses the related problem of arranging the four pieces to make a figureof area 63, as in Fig. 3.

Figure 6: AWGL.


David Singmaster 15

The oldest known version of this is an actual puzzle, dated c1900, [1], shown in[18] and there is a 1901 publication [5].

Loyd asserts he presented “this paradoxical problem” to the First AmericanChess Congress in 1858, but it is not clear if he means the area 65 version orthe area 63 version. Loyd would have been 17 at the time. If this is true, he isten years before any other appearance of the area 65 puzzle and about 42 yearsbefore any other appearance of the area 63 puzzle. I am dubious about this asLoyd did not claim this as his invention in other places where he was describinghis accomplishments. In 1928, Sam Loyd Jr. [11] describes the area 63 versionas something he had discovered, but makes no claims about the area 65 form,although he often claimed his father’s inventions as his own. For example, onp. 5, he says “My ”Missing Chinaman Puzzle” of 1896.

The first known publication of the 8×8 to 5×13, Fig. 2, is in 1868, in a Germanmathematical periodical, signed Schl. [14]. In 1938, Weaver [12] said the authorwas Otto Schlomilch, and this seems right as he was a co-editor of the journalat the time. In 1953, Coxeter [3] said it was V. Schlegel, but he apparently con-fused this with another article on the problem by Schlegel. Schlomilch doesn’tgive any explanation for this “teaser”, leaving it as a student exercise!

In 1886, a writer [13] says: “We suppose all the readers . . . know this old puzzle.”

By 1877 [4], it was recognised that the paradox is related to the fact that5 × 13 − 8 × 8 = 1 and that 5, 8, 13 are three consecutive Fibonacci numbers.Taking a smaller example based on the numbers 2, 3, 5 makes the trickery clear.

Figure 7: Nine to Ten.

One can also make a 5× 5 into a 3× 8, but then there is a loss of area from thesquare form.

However, there are other versions of vanishing area or object puzzles. Since



1900 several dozen have been devised and there are examples where both somearea and an object vanish!

Figure 8: CreditSqueeze.

This is a fairly common magician’s trick, taking a 7 × 7 square array of “£”signs and rearranging to get a 7× 7 square array of “£” signs and an extra unitsquare containing an extra “£” sign. My version, called “Credit Squeeze”, using“£”, is attributed to Howard Gower, but Michael Tanoff kindly obtained for mean American version, using $, called “It’s Magic DOLLAR DAZE”, producedby Abbott’s with no inventor named. Lennart Green uses a version of this inhis magic shows, but he manages to reassemble it three times, getting an extrapiece out each time! Needless to say, this involves further trickery. A version ofthis is available on the Internet.

But there are earlier examples. Gardner and others tracked the idea back toHooper [9] in 1783, as seen in Fig. 3. Here we have a 3 × 10 cut into fourpieces which make a 2× 6 and a 4× 5. However, Hooper’s first edition of 1774erroneously has a 3× 6 instead of a 2× 6 rectangle and notes there are now 38units of area. This was corrected in the second edition of 1783 and this versionoccurs fairly regularly in the century following Hooper.


David Singmaster 17

In 1989, I visited Leipzig and was reading Gaspar Schott [15] where I found adescription of a version due to the 16C architect Serlio [17]. I managed to findthe Serlio reference, which is in his famous treatise in five books on architecture.

Sebastiano Serlio (1475-1554) was born in Bologna and worked in Rome in1514-1527 with the architect Peruzzi. He went to Venice and began publishinghis Treatise, which appeared in five parts in 1537-1547. This was a practicalbook and greatly influential. He influenced Inigo Jones and Christopher Wren.Wren’s design of the Sheldonian Theatre in Oxford is based on Serlio’s drawingsof the Roman Theatre of Marcellus. Serlio also describes the “Chinese Lattice”method of spanning a roof using beams shorter than the width. This was studiedby Wallis, leading to a system of 25 linear equations for the Sheldonian roof. In1541, Francois I summoned Serlio to France and he founded the classical schoolof architecture in France. He designed the Chateau of Ancy-le-Franc and diedat Fontainebleu.

Figure 9: SerlioRot.

Fig. 9 comes from the 1982 Dover English edition, f. 12v. He is taking a 3× 10board and cuts it diagonally, then slides one piece by 3 to form an area 4 × 7with two bits sticking out, which he then trims away. He doesn’t notice thatthis implies that the two extra bits form a 1 × 3 rectangle and hence doesn’trealize the change in area implied.



Already in 1567, Pietro Cataneo [2] pointed out the mistake and what the cor-rect process would be.

I later found a discussion of this in Schwenter [16], citing another architect, sothis was well known in the 16−17C, but the knowledge disappeared despite thefact that Serlio’s book has been in print in Italian, Dutch, French and Englishsince that time and Schwenter was fairly well-known.

Since that time, I have found two other late 18C examples, possibly predatingHooper.

Charles Vyse’s The Tutor’s Guide [19], was a popular work, going through atleast 16 editions, during 1770-1821. The problem is: “A Lady has a DressingTable, each side of which is 27 Inches, but she is desirous to know how each Sideof the same may = 36 Inches, by having 4 foot of Plank, superficial Measure,joined to the same. The Plan in what Manner the Plank must be cut andapplied to the Table is required?” [The plank is one foot wide.] The solution isin: The Key to the Tutor’s Guide [19] (p. 358).

Figure 10: Vyse’s Solution.

She cuts the board into two 12”× 24” rectangles and cuts each rectangle alonga diagonal. By placing the diagonals of these pieces on the sides of her ta-ble, she makes a table 36” square. But the diagonals of these triangles are12

√5 = 26.83 . . .”.

Note that 272 + 12× 24 = 1305 while 362 = 1296. Vyse is clearly unaware thatarea has been lost. By dividing all lengths by 3, one gets a version where oneunit of area is lost. Note that 4, 8, 9 is almost a Pythagorean triple. I have not


David Singmaster 19

seen the first edition of this work, but the problem is likely to occur in the firstedition of 1770 and hence predates Hooper. Like Serlio, the author is unawarethat some area has vanished!

The 1778 edition of Ozanam by Montucla [12] has an improvement on Hooper.

Figure 11: Ozanam.

The image is Fig. 127, plate 16, p. 363 for prob. 21, pp. 302-303 in my copy ofthe 1790 reissue. This has 3×11 to 2×7 and 4×5. Here just one unit of area isgained, instead of two units as in Hooper. He remarks that M. Ligier probablymade some such mistake in showing 172 = 2× 122 and this is discussed furtheron the later page.

In conclusion, we have found that vanishing area puzzles are at least two hundredyears older than Gardner had found. We have also found a number of new formsof the puzzle. Who knows what may turn up as we continue to examine oldtexts? I think Martin would have enjoyed these results.

References

[1] AWGL (Paris). L’Echiquier Fantastique, c1900. Wooden puzzle of 8× 8 to5× 13 and to area 63. Described and shown in [10], p. 144.

[2] Pietro Cataneo. L’Architettura di Pietro Cataneo Senese. Aldus, Venice,pp. 164-165, 1567.

[3] H. S. M. Coxeter. “The golden section, phyllotaxis, and Wythoff’s game”.Scripta Math, 19, pp. 135-143, 1953.

[4] G. H. Darwin. “A geometrical puzzle”. Messenger of Mathematics, 6, p.87, 1877.

[5] Walter Dexter. “Some postcard puzzles”. Boy’s Own Paper, pp. 174-175,14 Dec 1901.



[6] Martin Gardner. “Geometrical Vanishes”. Chaps. 7-8 of Mathematics,Magic and Mystery. Dover, NY, 1956.

[7] Martin Gardner. “The author pays his annual visit to Dr. Matrix, thenumerologist”. Sci. Amer, pp. 138-145, Jan 1963. Reprinted in The Nu-merology of Dr. Matrix, Simon & Schuster, NY, 1967; in The IncredibleDr. Matrix, Simon & Schuster, NY, 1967; and in The Magic Numbers ofDr. Matrix, Prometheus, Buffalo, 1985. 1

[8] Martin Gardner. “Advertising premiums to beguile the mind: classicsby Sam Loyd, master puzzle-poser”. Sci. Amer, pp. 114-121, Nov 1971.Reprinted in Wheels, Life and Other Mathematical Amusements ; Freeman,Chap. 12, pp. 124-133, 1983.

[9] William Hooper. Rational Recreations, In which the Principles of Numbersand Natural Philosophy Are clearly and copiously elucidated, by a series ofEasy, Entertaining, Interesting Experiments. Among which are All thosecommonly performed with the cards. 4 vols., L. Davis et al., London, 1774;2nd ed., corrected, L. Davis et al., London, 1783-1782; 3rd ed., corrected,1787; 4th ed., corrected, B. Law et al., London, 1794. pp. 286-287: Recre-ation CVI - The geometric money. 2

[10] Sam Loyd, ed. by Sam Loyd Jr. Sam Loyd’s Cyclopedia of 5,000 Puzzles,Tricks and Conundrums. Lamb Publishing or Bigelow, NY, 1914 = Pinnacleor Corwin, NY, p. 323 (Get off the Earth); 288 & 378 (area 64 to areas 65and 63 versions), 1976. 3

[11] Sam Loyd Jr. Sam Loyd and His Puzzles An Autobiographical Review. Barse& Co., NY, pp. 19-20 & 90: A paradoxical puzzle, 1928.

[12] Jacques Ozanam. Recreations Mathematiques et Physiques, . . . (2 vols.),Jombert, Paris, 1694. 4

[13] Richard A. Proctor. Some puzzles. Knowledge 9, pp. 305-306, Aug 1886.

[14] Schl. [probably Otto Schlomilch]. “Ein geometrisches Paradoxon”. Z. Math.Phys. 13, 162, 1868.

[15] Gaspar Schott. Magia Universalis. Joh. Martin Schonwetter, Bamberg, Vol.3, pp. 704-708, 1677.

1As Chap 3 in each of these.2The image is taken from my copy of the 2nd ed.3This is a reprint of Loyd’s Our Puzzle Magazine, a quarterly which started in June 1907

and ran till 1908. From known issues, it appears that these problems would have appeared inOct 1908 and Jun 1908, but I don’t know if any copies of these issues exist.

4Numerous editions then appeared in Paris and Amsterdam, some in one volume; About1723, the work was revised into 4 volumes, sometimes described as 3 volumes and a supple-ment, published by Claude Jombert, Paris, 1723. ”The editor is said to be one Grandin.”.In 1778, Jean Etienne Montucla revised this, under the pseudonym M. de C. G. F. [i.e. M.de Chanla, geometre forezien], published by Claude Antoine Jombert, fils aıne, Paris, 1778, 4volumes. The author’s correct initials appear in the 1790 reissue] This is a considerable rework-ing of the earlier versions. In particular, the interesting material on conjuring and mechanicalpuzzles in Vol. IV has been omitted. The bibliography of Ozanam’s book is complicated. Ihave prepared a detailed 7 pp. version covering the 19 (or 20) French and 10 English editions,from 1694 to 1854, as well as 15 related versions - this is part of my The Bibliography of SomeRecreational Mathematics Books. The above is an extract.


David Singmaster 21

[16] Daniel Schwenter. Deliciæ Physico-Mathematicae. Oder Mathemat- undPhilosophische Erquickstunden, Jeremiæ Dumlers, Nuremberg, p. 451,1636. 5

[17] Sebastiano Serlio. Libro Primo d’Architettura, 1545. 6

[18] Jerry (= G. K.) Slocum & Jack Botermans. Puzzles Old & New – How toMake and Solve Them. Univ. of Washington Press, Seattle, p. 144, 1986.

[19] Charles Vyse. The Tutor’s Guide, 1771 (?). 7

[20] W. Weaver. “Lewis Carroll and a geometrical paradox”. Amer. Math.Monthly 45, p. 234-236, 1938.

5Probably edited for the press by Georg Philip Harsdorffer. Extended to three volumes byHarsdorffer in 1651 & 1653, with vol. 1 being a reprint of the 1636 volume.

6This is the first part of his Architettura, 5 books, Venice(?), 1537-1547, first publishedtogether in 1584. There are numerous editions in several languages, including a 1982 Doverreprint of the 1611 English edition.

7The problem is in the 10th ed., ed. by J. Warburton. S. Hamilton for G. G. and J.Robinson, London, 1799, prob. 8, p. 317. The solution, Fig. 10, is in: The Key to the Tutor’sGuide, 8th ed., G. & J. Robinson, p. 358, London, 1802.


Documents

Games and Puzzles - Ludus-Opuscularmm.ludus-opuscula.org/PDF_Files/Singmaster_Vanish_10_21(1_2014... · 12 Vanishing Area Puzzles Figure 3: Hooper. Let us look at the classic 8 8